And if so, which software do you use? I'm fascinated by storing things that I learn outside of my own brain. Ideally I'd love to create my own mini intranet of information, basically a personal wiki, that I can go to first before having to Google things, especially for my field of study.
I've looked into various Wiki packages but none quite seem to offer what I'm looking for (Mac, no need for a webserver, cross device functionality, math markup, searchability, tags)
If you do keep a personal knowledge repository, can you please describe your workflow for using and maintaining one for all the new knowledge you acquire?
I use adoc or markdown files in a personal git repo that I host on my own AWS instance. I use gitolite to manage access to the repo and even keep a version on my phone/tablet.
For years I used the FreeMind mind-mapping app, but wanted to move all of it to a server that I could use across devices once smart phones came along, and moved to flat text files in Dropbox that I can edit with any text editor. The other thing with FreeMind is that the UX for navigating mind-maps can get unwieldy when they get big. (maybe there are other apps that implement this UX better, not sure). I have also tried using wiki software ( like https://tiddlywiki.com/ ) and other tools that just didn't work as good for me as flat text files at the end of the day. I don't feel this is a solved problem and I would explore other solutions. I think something that has a timeline view and very good search tools as well as integrates with media libraries in the same search and timeline (like my lightroom and my phone's google photos libraries) would be a really cool way to see thoughts and images in the context of time going back to the day you wrote something, etc.
> The other thing with FreeMind is that the UX for navigating mind-maps can get unwieldy when they get big.
In what way? I have some pretty enormous ones and they seem fine so far. I'd like it if FreeMind were more actively maintained, but at least it still works basically fine on the latest version of Java.
I just remember expanding and collapsing things a lot to keep what I wanted in view. Clicking and scrolling -- perhaps I never got proficient in the keybindings so that could explain it.
Spacebar expands and collapses nodes. I think part of the problem is just that the default settings are bad. Changing these couple things makes FreeMind 10x better:
Preferences -> appearance
1. Check "edges start from 1 point at root node"
2. Uncheck "enter confirms by default"
I stick everything into Notes on macOS in an unstructured way and just use the search function to go back and find stuff. I also use Trello for more complicated things, but use that way less.
I used to use FreeMind [0], a mind-mapping tool that I really like, to keep track of projects, ideas, and things to follow up on. I eventually noticed that the pretty pictures could be represented, without loss of information, by a simple outline in a text file. Reading 'Getting Things Done' [1] got me to think in terms of a 'Current Task' list, which I now keep as a text file, backed up on Dropbox. Anything I need to do, or want to follow up on, or want to dream about in a practical way, gets jotted down in the file, with the most urgent matters at the top, Some really interesting stuff drifts down in the file as time passes.
As I do stuff, things move from the Current Tasks file to a Logbook file. Beyond 'accomplished todo's, that file gets stuffed with, for example, interesting quotes I run into, command lines I think I'll need again, and a record of what I've been up to. When I need to put together a status report, or figure out for myself where my time went, or recall how to do something intricate that I haven't done in six months, I turn to my Logbook.
With these two pieces, I feel like I've got a handle on what I want to do, and a record of what I've done. Now, if I just could find the time to follow through...
boxes of university notes losing their meaning year by year
and evernote, where I occasionally drop ideas, save snippets of code or commandline pipelines along with snipped articles from the web usually recipes or long form articles
I would benefit from doing more
More short term I write whatever is happening on especially long narrow ruled legal pads (8.5 x 14) which often still contain lots of information after a few years
I use Evernote. I have multiple notebook stacks for different topics (like each company I work on, ideas, general life notes, learning, research, etc.) and within each stack have a notebook dedicated to notes addressing topics as well (for example in the company stack have a notebook for sales notes, in the learning stack have a notebook for JS).
Evernote has worked well so far and me trying to switch to OneNote didn't work out as well. I'm still open to alternatives, but Evernote has kept my mind organised without a doubt.
I use and recommend Notion (https://notion.so) although I do work there, so there’s some bias.
It’s good for free-form notes, but also has databases and relations for more structured data. Also has top-of-the-like Markdown and HTML export to get your data out.
You can publish parts of your workspace publicly read-only, or share via email with guests.
Words cannot describe my enthusiasm for this tool - it's all I have been searching for.
I have tried many tools throughout the last couple years in an attempt to declutter my mind and manage my time better each one failed to meet my needs at some aspect: clumsy-UI, not online, not accesible by smartphone/tablet, lacking features and similar.
I can highly recommend this project to anyone skimming the comments - I think it'll suit most people's needs.
Just searched to find the comment to upvote. Notion changed my life! It took over my life. It is my everything. My life is in Notion now. It is my repository of life. I am not exaggerating. I do not work there. It is SUCH a quality product in every way. You have features that I only even think about after using it every day for 8 months.
I've had problems finding a note taking suite that I could really buy into. There are a few I've used, they all have some pros and cons, but nothing stands out right now as the clear winner:
* Tomoby:
- pros: fast, good search, really liked local editing, integrates with quick launchers like Gnome DO and Albert which I found essential for quick references, especially programming and Linux commands which, when I want them, I want them NOW. type: "<ctrl>+<space>grep examples" and have a reference to my most frequently used patters in 1 sec is SO helpful.
- cons: sync/mobile solutions are hacky, no longer maintained, Ubuntu 19.04 removed it, not encrypted, Copy & paste to/from other apps (Thunderbird) doesn't work like you'd expect, formatting is not preserved.
* Evernote: meh...used it for a couple years. Liked multi-device support. Their web client wasn't that great and they didn't have a Linux client. I paid for it and put maybe 100 notes in it, but eventually stopped using it and converted back to the free version.
* Dropbox Paper:
- pros: nice editor, have the basics right, integrated with Dropbox which I already use, so no additional cost, good mobile client, really easy to share notes with groups of people (family) or one-offs to others with a link.
- cons: some bugs and bad UX still, no encryption (but Dropbox already has most of my files so I've already sold out a bit to them I guess), loading a browser tab is just slower than the Tomboy UX described above...I find I like using Paper for longer docs or things I'll access on mobile, but it's just not a good replacement for those quick notes that when I want them, I want them NOW!
I've tried a number of other ones over the years, never really found a sweet spot. I actually think something like Tomboy is really the best note taking app, but for it to really thrive, it would need better options for mobile and note sync than Tomboy ever had. Adding end-to-end encryption and sharing would also be great.
- Excerpts (for summaries and excerpts from books and articles)
- Snippets (for code snippets, formulas, etc)
- Reports (for results of research on a subject, e.g. learning new API)
- Papers (copies of research papers)
Dividing notes in such a way also helps to review them with varying frequency. For example, snippets get outdated often (e.g. workarounds for bugs or polyfills) so I have to go over them more often. Reports get updated as I learn new things about a particular subject that has its report. Excerpts are updated only if I decide to re-read the material they cover and extend them. Papers last as long as I'm interested in subject.
I use google colab (https://colab.research.google.com/notebooks/)for things like work notes on setting things up, any new tech I am learning. The best thing is it support code(python).
Google keep(http://keep.google.com) for noting down any interesting thoughts/ideas/lesson I learn through out the day. I usually create a card per month to keep it organised.
Really interesting idea to use Colab. I assume that you occasionally download the notebooks for local backup? Also, are downloaded notebooks useable with Jupiter notebook?
Yes I do but I wish I didn't keep it so haphazard.
By haphazard I mean to focus not only on acquisition but reflection / contemplation: Collecting thoughts is one thing, putting them together into a cohesive and coherent body of knowledge is another. So, that's what I'm working on.
pmwiki for now. Workflow is personal domain, putting browser or window on one side of the screen and what I'm reviewing, or what I'm thinking / text editor (sublime text) of the other side. Putting into a DIY solution based on flask simply because that's flask practice.
I also like pens and paper and whiteboards. So take pictures and use Powerpoint's 'remove background' feature to have a liberal amount of pictures.
A bit rambling, but you asked and asking is usually best! So.. :D !
I've had troubles coming to a good software solution or workflow either.
Software wise, I've tryed a few things and settled with MediaWiki and some extensions added. Most notably Semantic MediaWiki (adds structured data storage and querying), Semantic Forms (structured forms) and a Bootstrap based skin. But only the technical setup won't get you far in a wiki, so I've also created a lightweight, flexible "ontology" / structure for knowledge management that was fitting my own needs.
This was not too easy to set up (Finding a structure / ontology / data and query workflow, but also the implementation using the wiki itself), but now I'm quite happy with it as a platform. I really wish there would be a lower barrier to entry for people here. If you're interested, please PM me and I can show you some examples.
As for the workflow of personal knowledge management, I'd be really interested what works for other people as well and I'm still figuring this out.
My current approach is using different phases. It's OK for me to read a book and not note anything or only do kindle notes. When reading articles on the net, I usually bookmark more interesting articles using getpocket. I do notes on paper when I don't want to be distracted or influenced by a computer.
For this year, my resolution is to acquire less new knowledge, but to take some time and to rehearse, summarize and repeat what I've found interesting and worth my time before. What sticks here should go into the wiki.
I want to keep my knowledge base very concise and opinionated. There's no need to create a second wikipedia.org. I'll only add something in the wiki when:
* Some idea is truly remarkable on the first spot
* I go over my old notes see what was the essentials of it. Those are worth putting into a wiki.
* When I'm doing some dedicated, structured research around a topic. Then I also try to get the ontology and structure right first.
* Code Snippets that I repeatedly use
Both public and private. Public, a static site made from markdown files : https://raymii.org. Almost 400 articles now.
private, a folder called "notes", with almost a thousand short text files. Workflow for it? Use grep to find what I want. I had a system of folders per topic but that became a mess. Grep is good enough. The notes folder is synced with synching to multiple systems and my phone. On the phone I have termux, which has grep and markor as text editor.
You can check out Anki[1]. Apart from the initial setup, you won't have to worry about taking care of the web server.
I create separate decks for different fields and then only if necessary do I create subdecks. I also tag all the cards that I create and make sure to spend a lot of time to ensure that it's made and tagged properly.
So, whenever I learn something new and I believe at that point of time that it's worth remembering, I create a card for that.
I use the AnkiDroid app on my phone. And it's a pain to quickly make cards on the phone. You will get better and faster at creating cloze deletions after creating a few cards on a phone.
That's why I only use AnkiDroid on phone to revise (and seldom create cards) if accessing a laptop is a bigger pain.
Deft in Emacs. [1] I write my notes in Markdown or Org-mode, and Deft lets me search them all really fast. For cross-file references, I use Org’s delightful file hyperlink scheme. For export, I use Pandoc. I’ve got a project that I haven’t quite figured out how to get on home brew (I’m new at this) that takes a file in <format> and lets me read it like a Man page. [2]
Wow. Rifle looks neat. Thanks for mentioning that!
Deft will scan any plain-text file, according to the set of file endings you give it, and it will put your cursor at the point in the buffer where one of the search words occurs, but it won’t automatically extract the ending, AFAIK.
I should mention that deft does fuzzy matching, or at least words don’t have to be contiguous.
Lately I've just been using a directory full of markdown files. I store my personal collection in Dropbox but it would be just as easy to use git or any other tool. I've created an alias to automatically open that directory in my favorite editing tool when I type "notes".
I also manually maintain an index.md file to categorize things, but it's not really necessary when you can just use grep or your editors search feature to find what you need.
I do the exact same, down to the "notes" alias. I find the downside to this approach is it doesn't work well on mobile - the Dropbox app can edit files but there's no search and it all feels a bit clunky.
I'm usually in a terminal, so I use vimwiki (https://vimwiki.github.io/) - it supports markdown, adding pages is trivial, and vim tags or CLI tools like grep makes it easily searchable.
It lives in a private NextCloud instance, so everything's available from my laptop or phone (although editing markdown on a phone is no fun).
Edit - this is great for detailed note-taking; I also use a custom command-line tool to capture ad hoc notes, ideas, and suggestions without opening vim. They go into a YAML file right now; tooling in process to pull them directly into the wiki, so everything's in one place.
Yes. I write things in Tree Notation using a simple file backed plain text database called Treebase. It allows me to easily jot information down using nearly any program and maintain a strongly typed, color highlighted, queryable database.
I have a few different databases now, the biggest in the millions of rows. A workflow for one of the databases, an ontology of human medical measurements, is as follows: 1) I come across a new term, 2) (if afk I first email it to myself) I simply do a “touch term.hipo” at the shell, or use vim, sublime, or another handy tool. Then I add data in a strongly typed but 0-syntax tree language I created called “hipo”. If I’m in sublime or another editor I get live autocomplete and syntax highlighting 3) I use git to version everything 4) I have a simple DB server which reads the files and allows me to analyze using any tool, from R studio to Vega, etc.
It's fast and has good search, a good editor, and uses simple hierarchical organization + tagging. I like that it gets out of my way and does a good job of handling everything from code snippets to freeform thoughts to recipes.
BookStack is best for things that 'look like' documents. Mobile editing could also be better - you have to click a maximize button to get a usable view of the editor and the editor UI still takes up a lot of space. It is web based, but it is very lightweight and I have found it very easy to host.
BookStack developer here. Thank you for sharing the project. If you have not yet updated to v0.26 it might be worth doing so as, upon many other UI changes, I focused on trying to improve the mobile UX [1]. Still some issues in iOS to resolve though.
Notable is a great free markdown editor on GitHub. You can import regular markdown or Evernote files. All the files are stored in one directory and are grouped by tags. The tags can be hierarchical if you’re looking for that. Most of the time though the regular file search does great.
I find myself making so many small notes of how I did things at work or with AWS, zsh etc.
It’s a really powerful system while simple at the same time.
You would have to pry https://workflowy.com from my cold, dead hands. It's the best outliner I've ever used and they actively resist adding features that would only please a subset of users. It offers excellent cross-device synchronization and works nearly as well on a phone as it does in a browser or desktop app context. You can attach notes and share at the node level, either via read-only link or collaboration invite which allows others to graft your shared hierarchy to theirs. It has well-considered keyboard shortcuts.
Most importantly, the search is intelligent and works instantly... even across my 100,000+ nodes.
My entire life is in Workflowy. I am thrilled to pay for my annual subscription, because the value I receive is outrageous.
Workflowy looks quite interesting. Could it be used to store in-depth knowledge? My quick play with it on their site looks like it's more of an outline tool, but how would it go with say storing the information that you might find on a Wikipedia page?
I do think a truly useful knowledge repository will fundamentally need to be a graph structure. This roughly reflects the neural connections in the human mind -- knowledge itself is a graph structure.
I am using Joplin for all my knowledge articles and notes. Available for all main desktop and mobile OS's. Great Evernote replacement with good Webclipper. Notes can be edited in Markdown. Syncing to Nextcloud, Dropbox, Onedrive, Webdav. And fully open-source. Currently a few 1000 notes, searchable or browsable via notebooks or tags.
I wrote my wiki server [1], as I was missing a proper freeware at that time. I run one instance at home for my private stuff and one on my office computer for confidential infos related to work.
The wiki stores pages in simple text files that are easy to backup, it automatically calculates birthdays of my contacts (stored with semantic annotation in an XML style), and provides an Android App to synchronize all pages on a mobile phone. Many wiki pages are organized hiararchically, but most times I use the full text search.
Technically it runs with Java on any OS, and has a built-in web server.
Twitter Favs and Safari Reading List (iOS and Mac) • to quickly capture items I’ll want to revisit in the future. For instance, I added this HN discussion to my Reading List, so that I can later review the other suggestions and try them out.
Drafts (iOS and Mac) • quick thoughts I want to bang out before they dissolve. Drafts is focussed on being the fastest zero-to-blank-page experience on iOS, which makes it very handy for rapid spur of the moment universal capture.
Quiver (Mac) • I periodically review the above two, and catalogue them here for long term storage. There’s no structure to review, this is just my brain’s external HD.
I am currently experimenting with Notion, really well done knowledge orgination, but not open-source and paid... so long term might be hard. Anyone has any experience with Notion?
I started with the augmentingcognition.com spaced repetition model, and then decided to augment it with a catalog of everything I read(1). For my own original thoughts, I write them down in a few paragraphs and then treat them like any other reading material.
At some point, I ran across the idea that a library card catalog is really just a combined topical index for the entire collection, with each entry being a card. That seemed like a decent model, and I’ve been happy with it so far.
My system is currently a physical card index because I needed the flexibility of pen & paper to figure out how I wanted it to work, but I’ve been slowly computerizing it as I figure things out- you can certainly do something similar with flat text files, spreadsheets, or any other general purpose document software.
So now, I have three interacting systems: a citation list that records where I can find things to read them again, a combined topical index of everything I’ve read, and a flash card deck of things I want to remember without having to look them up.
When I run across an interesting article on the web, for instance, I save a PDF of it to Dropbox for safe keeping and print a copy to read away from my computer. As I read, I’ll make margin notes and have two different colors of highlighter: one to note index headings and another to mark passages to be turned into flash cards.
After my morning flash card review, I go through the material I read the previous day and turn all of the notes and highlights into cards(2). Once everything has been moved into the indexing system, I have no further need for the physical copy, so it goes to the recycler.
(1) Of sufficient length; generally 3 or more printed pages.
(2) Initially by hand, now with the aid of a custom piece of software
I have something called "Diensttagebuch", which is awkward to translate ("Duty diary"?), it's a Jekyll blog (currently at https://www.pchr8.net/d/dtb/) for things I know I'll have to google later. I have two bash scripts for creating-or-opening a new page and for deploying (upload to server + commit to git). I use https://github.com/rking/ag.vim for searching though the Markdown files on my local machine.
For more general knowledge, Semantic Mediawiki (http://pchr8.net/f/), but it's basically a big link database now. I can sort it by rating / complexity / to read / domains / keywords and found it quite flexible.
How long would you like to be able to access this content?
If it is for the long haul, plain text is the only way to go (org-mode, etc).
Any proprietary commercial solution will fade away in a handful of years. At best, you might be able to extract the content before they shut down but structure and metadata may be lost.
Open source self-hosted solutions will last much longer but even in that case it might get to a point the source has been abandoned so long that it becomes difficult to keep going.
Only plain text will survive. I have plenty of notes in a more or less organized directory tree, dating back to the late 80s.
I am wondering what information from 1988 might still be of help for you now, but I would suggest that xml and json or similar would still last about as well as plain text, but would enable you to do more interesting things and store more useful documents (say a complete contact list).
XML or JSON are also plain text. And Org has excellent support for storing contacts.
Furthermore, a personal knowledge repository may contain information that is pretty much timeless. Say you read Rudin to learn analysis in 1988 and took some notes. Why would that be outdated?
Some academics build huge perosnal knowledge repositories that they keep growing during their careers. These will often span longer periods than 1988-2019. See e.g. Luhmann's Zettelkasten.
The advantage of Org is that it is very readable, as it descends from an outliner, and has excellent tooling. There are options for including tables, equations, bibliographies, links to many datatypes, attachments, extensive date support, etc. Lots of export tools, interpreters to show views of your document, and many other things I have not touched upon.
Yet it is quite simple hierarchical plain text. So it will remain accessible in the future. Aside, Emacs has been around for very long, and will probably remain around much longer due to the Lindy Effect.
I use Notion.so and my second brain. I used to store all this info in markdown files spread across many folders. The most important tool that Notion has for me is a quick search over all this data.
The most important for me is quick storage and retrieval, so everything must be keyboard based. To make a note, I press a global hotkey, an editor window pops up, I write the note then close the editor with a hotkey.
It makes it very seamless to make new notes.
And for retrieval I press another global hotkey, a prompt pops up, I start typing and it immediately lists my matching notes (the search query can find the typed words in any order of course), I select from the list with enter and the note is displayed, and can be edited too if needed.
So finding something takes usually at most 5 seconds (pressing the hotkey, typing, selecting). A mouse-based system could not beat this efficency.
Above I only described the general concept which can be implemented with various tools.
I personally use Autohotkey on Windows for the global hotkeys which when pressed activate my running emacs for new note or search (I use orgmode in emacs for notes). And for searching I use the org-rifle package.
I saw somewhere that Woody Allen keeps a bin full of pieces of paper that have 1-2 sentences of a possible plot of a movie, just throws another one in when there's an idea.
I have been using personal knowledge repositories for more than 20 years now, and if there is one lesson I have learned: never, never ever use any proprietary, platform specific - or worse, web-based - solution, no matter how good it is. Plain text is the only way to go.
My favorite tools in the past were Tinderbox from Eastgate and Curio from Zengobi, but both were Mac only and I had to leave them behind when I switched to Linux a couple of years ago. I collected some experience with Scrivener and used PMWiki a while, but the first is not really supported on Linux and the second wasn't somehow responsive enough for my needs. ZIM wiki on Linux is actually quite nice, but when I discovered orgmode it hit just right and I never looked back for other tools.
My org-file organisation: I have one super-big master org-file in which I collect just about everything regarding software development, technology in general, life lessons and stuff. For customer projects I generally create a dedicated file in which I not only collect knowledge regarding that project, but also do the project management with tasks, documentation and stuff. If one of the topics in the master org-file grows really big (this is for example the case, when I start to use a new programming language full scale for productive purposes) I will migrate that node to its own file. I also keep most of my bookmarks as links in org-files in a proper context, and I like to use links to files in the local filesystem (like PDF-manuals) within my org-files, although this requires some discipline in the organisation of the local folder hierarchy.
The only other tool that I started to use besides org-mode, just a couple of weeks ago is Anki, for stuff that I want to learn by heart.
I went through many software solutions and ended up with Org Mode as well. There may be some clunkiness at the beginning and it definitely lacks the glamour of proprietary solutions such as Notion. But once you get a grip on Org Mode, you will be amazed how deep the rabbit hole goes. ;)
There is a great Org Mode tutorial [1] on YouTube by Rainer König.
The only « problem » is that you need to learn Emacs. I recommend Emacs Doom [3]
Also, I'm experimenting with making my Org Mode notes public. Here's a humble beginning [2]. I only publish them once they are cleaned, so it's just a tiny bit for now.
I've been looking for something similar to a browser history, but it actually snapshots the text/images for each page so that I can search it. Ideally it would be OSS, self-hosted and cross platform so that it can feed a repo from desktop web and phone.
Seems like something as simple as a firefox add-in would be useful.
Memex by World Brain does this pretty well. Open source and stores all your browsing with great search. Can tag and highlight pages, as well as back up to cloud services
For years I used Evernote but give it up a few years ago.
Now I use a combination of org mode, PDFs and miscellaneous document types replicated in both Dropbox and Google Drive, and on macOS the system search finds things. Dropbox search only hits file names, Drive search is more useful.
In addition to my own notes, I collect PDFs of good papers, copies of books I have bought, etc.
EDIT: forgot about Google Keep and Apple Notes - sometimes use for specific things like vacation travel notes
I've found markdown + VS Code/whatever editor perfect for it. You get full directory traversal with its system and get to view the files in a formatted way (even if the formatting marks are still there) while being able to edit on a whim. I'm considering git for it as well. You can also cloud sync your folder to other devices on any number of providers.
For me, I use Orgmode, plus a mess of PDFs in folders on Google Drive. I've been looking into both Polar and Emacs as potential replacements for document organization.
Right now I have a single folder that I save new things to, and I have a recurring task to go back and file documents regularly. It's not perfect but it works. I don't often reference my pile of PDFs, though; I might use it more with tagging options, which is why I'm looking at replacements.
I used to use Evernote. There were some things I didn't like about it, although it's been so long that I couldn't tell you what they were anymore. It might be a good system now.
Yeah. I am currently using org-mode (plus org-wiki, some bespoke elisp and a few shell scripts) for my own knowledge base, but Oscean/the Nataniev ecosystem in general has impressed me.
plus Zotero for papers/PDFs, plus a pile'o'files stashed using the Universal Decimal Classification for its folder tree, plus assorted odds and ends.
In fact, I spent this week messing around with some of its concepts - Indental plus Runic plus a lightweight markup language loosely based on txt2tags + sexps and which should be easy to render even outside the browser.
The idea is to move away from documents (which org-mode is very good at), and towards smaller, rich chunks of information (which org-mode can't handle that well - I hacked things up with properties, but that namespace is flat).
> The idea is to move away from documents (which org-mode is very good at), and towards smaller, rich chunks of information (which org-mode can't handle that well - I hacked things up with properties, but that namespace is flat).
I've noticed this weakness in org-mode as well. It's really a fantastic tool, but it still has a few gaps here or there. I'm optimistic about its continued development though.
It's good to hear that you liked the Oscean/Nataniev tools; I haven't tried them and it could easily have been all style no substance. They certainly sound impressive. From what I've read, I have a feeling there might be something worthwhile there. And XXIIVV itself looks amazing, although I know that's just one implementation.
> I'm optimistic about its continued development though.
Org is amazing, yes, but I'm not too optimistic about this - even if some of its more noticeable warts got fixed (a big one, for me, is the fact that you cannot nest quote blocks), it would still represent a tree of nodes, and trees are somewhat limiting where it comes to knowledge representation.
You can organize information in a tree, but it doesn't neatly fit that model, and just deciding where to put each snippet you captured can be a huge hassle.
> It's good to hear that you liked the Oscean/Nataniev tools; I haven't tried them and it could easily have been all style no substance. They certainly sound impressive. From what I've read, I have a feeling there might be something worthwhile there. And XXIIVV itself looks amazing, although I know that's just one implementation.
I'm not too sure, either, but I'm going to take a stab at implementing something Oscean-like. I want to see where this approach can take me.
A bunch of markdown/text files in a git repo, hosted on github/gitlab/gitea (github in my case).
This way I can edit the files online on github.com, on my desktop and push/pull the changes. On my iOS device I use http://workingcopyapp.com - I can edit and create files as well as push and pull changes.
On my Desktop I can use any tool I want to search and edit the files, on github I have full text search and working copy also features full text search.
Most importantly: I own my data and they don't up in some database I need to worry about.
I have give a lot of thought to this. I bookmark a lot but its always hard to find information later on when you need it.
I came up with my own mental model “Make Knowledge Actionable”. With that, I find it is important to align knowledge you are saving with your goals. This means that someday I might need it knowledge should be kept in a different spot.
If I have some time, I have thought to build a system that lets me find information via mind map and different categories of abstraction.
Yes, I built an open source tool called Dnote [0]. Basically, I can write down my microlessons as I learn them without leaving the terminal or browser. And it sends me a weekly digest to automate spaced repetition.
I used to use Evernote to write down my knowledge but I quickly found that I never look back at my notes, and quickly lost grasp of my knowledge. That's why I built the automated email digest.
My workflow is simple: when I learn something, I execute a command line program to write it down without leaving the terminal when I am coding. Every weekend, I receive a digest of what I learned in the past. This way, I can successfully maintain all the new knowledge I acquire.
Similarly to your requirements, I often search through my knowledge base in the terminal before doing a web search. Doing so saves me time because my knowledge base is already written in my own words that I understand well.
Notice the timestamps (the alias opened vi and put a timestamp).
I also use TaskWarrior and have an alias to commit and push the changes to the files to a private repo.
I should collect dispersed files on my disk that are the result of reading articles and books (I open a file in the same directory of a paper or a book and start writing thoughts at a coffee shop, which I send to coworkers, especially for papers).
Bottom line, now I use MKdocs with GitLab pages (for now). It works nicely. The alias is good for quickly jotting down to a file and saving, then it's cleaned and edited and incorporated in the proper file in the knowledge base.
Yes, I use checkvist. ANYTHING I find mildly interesting goes into this list. This is great for when I come to something that I need to research or want to start going deep into; when this happens I've already accumulated dozens of links and bits of data to follow up with. As of the time of writing this, I have over 6000 unique links or bits of info. It's all organized by fairly arbitrary naming that I sometimes change.
E.g. my top levels are:
- server administration
- networking
- programming
- telephony
- science
- mathematics
- culture
- management
- job search
Last year, I went through this whole gamut of the usual suspects emacs(org-mode), plain text, markdown, tiddlywiki/ mediawiki/bunch of other wikis, google keep, one note, notational velocity, anki, evernote, LaTeX. One after the other, they all crumbled. I tried them all and failed. I got to a point of calm reflection based on my experience with each one. I realized what was most important for me. I eventually settled on a combination of email + git + google calendar. I use a separate email just for my notes. Every new thought, todo, bookmarks, quotes, images of doodles, event, information about people becomes a new email. Any update on the same one becomes a reply to the last email on that subject. Any topic or area that gathers enough content or insight that’s worth remembering goes to my git repo’s wiki page. If there’s enough content/files/code, it gets a directory. Common patterns, get linked via their respective wiki pages. I create collections of related topics as separate wiki pages with links back to the source.
This setup has worked well for me for over 6 months now. I do not worry about email / git becoming obsolete. Or how I would move my email/git data when I need to switch vendors. I can access it via any device I own or through someone else’s. Stuff I want to share with people are hosted via git pages or sent via email. I have a plethora of client applications to choose from. And above all... it just works.
I can give out more details but I feel anyone’s personal information needs can be met/exceeded with a combination of email + git + calendar.
This is a fascinating approach and would actually solve a lot of the problems I've had with trying this before. I would love to hear more if you had anything additional to add because this method has become the top of my list after trying several software packages mentioned thus far.
I'm happy this appears to solve your problems with other approaches. I took out a few minor but important details. use a separate email client for this on all devices. Use conversation/threaded view. Have a way to send an email in under 5 sec. I use iOS shortcuts to quickly send an an email with just the subject. Review the list of emails almost everyday. Have folders in your email and move them as necessary. Keep it under 5-7 folders. Ideas,Bookmarks,Done,Projects. Inbox is your list of ongoing high priority tasks. Stay away from adding petty tasks(use a paper/pen for this). Trash task emails once you're done. If its worth noting, it goes to git.
Hi. I also thought of email when I was thinking about this problem recently. How do you get information from your email thread to the wiki? And is this a github wiki?
Email is my todo/inbox. Its disposable. Wiki is gitlab wiki. It is my index for the information I've collected.
I do not move too much information from email to wiki. I'm quite mindful of it actually. Wiki gets written and re-written, cross linked and edited. I keep it consistent. Helps quick lookups. I very frequently use markdown files in topic specific directories. Helps store information gathered from deep dives on those topics. Markdown file names are also consistent - YYMMDD-topic-detail.md. Helps sorting by date.
I use the wiki mostly for the free hosting and the links. I could very well use a static site generator when there's enough content in a topic's directory to be shared as a mini website. A link to the wiki on the home screen comes quite handy.
I think a public GitHub in markdown format is a wise choice. Firstly markdown can look nice but will always be readable using cat (not what cats for but leaving that to one side...), so there is the longevity factor. Using github means you’ll probably have a local copy, and you can subject the repo to your normal backup procedures. But you also have another backup in Github itself. Finally making it public will encourage you to curate it and keep it up to date and not blasé delete old stuff (and git protects you from that too because a force push is required to really delete something). Most importantly it’s easy - open any editor and start writing, then commit ‘n’ push when done.
I use DevonThink. It's an incredible piece of software.
I have a database for 'research' where items are stored according to the Dewey Decimal categories (no particular reason for choosing this other than it's well organised and well documented).
I then have a separate 'memex' database which contains my journal and other miscellanies.
DevonThink has a rudimentary AI system which attempts to link documents by context. I find it quite useful.
zettelkasten.de has lots of good reading (thanks, Sascha and Christian!), but look to the other sources I've listed to get a more thorough idea of what a zettelkasten is (I call mine a thinking-box) and how to use one.
Yes, I write on little journals, I have almost 6 yrs doing it and I have a little box full of them, It isn't organized but it's always nice to look through when I'm searching something.
I keep a personal library (of physical books) for this reason. People often underestimate the value of spacializing knowledge. I write in the margins, mark them up, and leave scraps of paper with notes. Then I put them on shelves in broad categories - for me that’s things like architecture, design, technical, photography, philosophy. The result is a system thats’s easy to retrieve from, and importantly, easy to browse. The spines become symbols for collections of thought. I have yet to come across a piece of software that more effectively facilitates drawing inferences and connections from deep material. IMHO everything else can be found on Google.
LOL! Yeah, in my browser bookmarks! It...Is...A... MESS! I basically have it in typical outline form: I.A.1.a.i.,etc., without the I.A...headers, but rather, titles. I noticed the other day that I have 3 different Linux Main chapters, 4 Tutorials Main chapters...and suddenly, an epiphany. It reminded me of my brain: a lot of things in there, but a real pain to try and find. And those mind- mapping apps? Why? They don't have any semblance of order either! They looked like the web from a spider whose all 4 pairs of eyes are crossed. I am trying to brainstorm, but then my ears fill up with water from the storm, and those tiny lightning bolts are painful! I will figure something out. Oooh! Time to go play in traffic! Momma says it's good for building up my reflexes! K, bye!
~MJC
I use Google Keep and add all my personal learnings under a common label - <name> Wiki. Google Keep has excellent search capabilities too so easy to find what you're looking for.
I've looked into various Wiki packages but none quite seem to offer what I'm looking for (Mac, no need for a webserver, cross device functionality, math markup, searchability, tags)
If you do keep a personal knowledge repository, can you please describe your workflow for using and maintaining one for all the new knowledge you acquire?